How do consumers in developed countries value the ...annecelia.disdier.free.fr/FP.pdf · How do...

35
How do consumers in developed countries value the environment and workers’ social rights in developing countries? Anne-Célia Disdier Stéphan Marette Paper published in Food Policy 37 (2012): 1–11. Abstract This paper uses a lab experiment to investigate developed countries’ consumer valuations of characteristics linked to the environment and workers’ social rights in developing countries. It focuses on seafood products and distinguishes between regular, environmentally friendly and fair trade varieties. Consumer valuations are elicited with a multiple price list. Results show that environmental and social labels have similar effects on participant willingness-to-pay when they are first presented. Using welfare variation coming from the labels, we also show that the absence of negative information linked to the regular variety may lead to an underestimation of the value of information associated with the label. Keywords: label, willingness-to-pay, multiple price list, value of information, seafood. JEL Classification: C9, D8, J8, Q5. Financial support received by the “AgFoodTrade - New Issues in Agricultural, Food and Bioenergy Trade” (Grant Agreement no.212036) research project, funded by the European Commission, is gratefully acknowledged. The views expressed in this paper are the sole responsibility of the authors and do not necessarily reflect those of the Commission. Corresponding author: Paris School of Economics-INRA, 48 boulevard Jourdan, 75014 Paris, France. Email: [email protected] INRA, UMR Economie Publique INRA-AgroParisTech, Avenue Lucien Brétignières, 78850 Thiverval-Grignon, France. Email: [email protected]

Transcript of How do consumers in developed countries value the ...annecelia.disdier.free.fr/FP.pdf · How do...

How do consumers in developed countries value the environment and

workers’ social rights in developing countries?

Anne-Célia Disdier Stéphan Marette

Paper published in Food Policy 37 (2012): 1–11.

Abstract

This paper uses a lab experiment to investigate developed countries’ consumer valuations of characteristics linked to the environment and workers’ social rights in developing countries. It focuses on seafood products and distinguishes between regular, environmentally friendly and fair trade varieties. Consumer valuations are elicited with a multiple price list. Results show that environmental and social labels have similar effects on participant willingness-to-pay when they are first presented. Using welfare variation coming from the labels, we also show that the absence of negative information linked to the regular variety may lead to an underestimation of the value of information associated with the label.

Keywords: label, willingness-to-pay, multiple price list, value of information, seafood.

JEL Classification: C9, D8, J8, Q5.

Financial support received by the “AgFoodTrade - New Issues in Agricultural, Food and Bioenergy Trade” (Grant Agreement no.212036) research project, funded by the European Commission, is gratefully acknowledged. The views expressed in this paper are the sole responsibility of the authors and do not necessarily reflect those of the Commission. Corresponding author: Paris School of Economics-INRA, 48 boulevard Jourdan, 75014 Paris, France. Email: [email protected] INRA, UMR Economie Publique INRA-AgroParisTech, Avenue Lucien Brétignières, 78850 Thiverval-Grignon, France. Email: [email protected]

1. Introduction

Environmentally friendly and fair trade labeled products have emerged over the last two decades

and their markets have expanded very rapidly. By enhancing production and export capacities,

such products may contribute to the emergence of developing countries on the world market and

to their economic growth. Some factors may however restrict their sales. At the macro-level,

multilateral trade negotiations at the World Trade Organization (WTO), while acknowledging

the important role played by such labels, do not make them mandatory. For instance, under WTO

rules, countries cannot ban products based on the way they are produced (environment pollution,

indecent working conditions, etc.). At the micro-level, consumers may not really trust such labels

or are not willing to pay a premium for such products.

In this paper, we investigate developed countries’ consumer valuations of characteristics

linked to the environment and workers’ social rights in developing countries. Developed

countries are indeed the main market for labeled products coming from developing countries. We

use a lab experiment conducted in France in 2009 to evaluate the impact of information about

environmental and social characteristics of products on consumer choice. We distinguish

between regular and labeled varieties. Production of regular varieties may imply pollutions

and/or may not respect decent social rights, while labeled varieties goods are made according to

certain environmental or/and social production standards. Successive positive or negative

information is delivered to participants. Our experiment focuses on shrimps. Several

environmental and social issues affect their production in developing countries.

World shrimp production has grown rapidly during the last two decades. However, this

boom has come at some cost. First, there are health costs as shrimps often contain bacteria

(e.g. salmonella) or pesticide, drug and antibiotic residues. There are also concerns related to the

environment with the destruction of mangroves and the depletion of rivers and groundwater

sources for maintaining oxygen levels in farms. Other concerns deal with the illegal use of areas

3

for shrimp aquaculture and corruption of local authorities, as well as bad working conditions

(World Bank, 2001). These costs are likely to play an increasing role in swaying both

consumers’ choices and international trade. Environmentally friendly shrimps have recently

emerged. However, they represent less than 1% of world production (Hervieu, 2009).

This paper makes an important contribution to the experimental literature on labels. This

literature shows that a significant proportion of consumers are willing to pay substantial

premiums for environmentally friendly products (Blend and Van Ravenswaay, 1999; Bougherara

and Combris, 2009; Loureiro et al., 2001; Nimon and Beghin, 1999; Wessells et al., 1999) or fair

trade products (Arnot et al., 2006; Loureiro and Lotade, 2005). However, there is still

disagreement on whether or not labeling should combine different characteristics to increase

WTP and to favor products from developing countries. Loureiro and Lotade (2005) compare

WTP for coffee with fair trade, environmentally friendly or organic labels and Bernard and

Bernard (2009) compare WTP for milk with conventional, organic, rBST-free or no-antibiotic

characteristics. These papers offer a ranking of positive premiums for these new characteristics

signaled by a label and suggest that consumers are very receptive to one characteristic. Our paper

goes one step further since we show that the ordering of information is important. Participants

are receptive to the first-detailed characteristic presented in the experiment. However, we also

highlight a fast-diminishing interest in the added characteristic presented in second position.

The second contribution of our paper is to provide a complete estimation of the value of

information associated with labels and defined by consumer surplus variations derived from the

experimental results. While previous papers have determined the value of information through

experimental auctions or surveys, we attempt to do so with a choice procedure based on a

multiple price list. Furthermore, the estimation includes all negative and positive information for

a characteristic (social or environmental) that delineates the regular and the new labeled

varieties. In many existing studies (Huffman et al., 2003, 2007; Lusk et al., 2005; Lusk and

4

Marette, 2010; Rousu et al., 2004, 2007; Rousu and Lusk, 2009), 1 all information revealed in the

experiment concerns only the newly introduced variety and not the regular/conventional variety.

Our paper shows that the omission of the negative (or positive) information linked to the

regular/conventional variety may bias the estimation of information value. More precisely, this

omission leads to an underestimation of the value of information associated with the label since

the knowledge about the regular variety is still imperfect.

The next section describes the experiment. Results are reported in section 3. Section 4

provides econometric estimations of the determinants of consumer willingness-to-pay. Section 5

studies the value of information and consumer welfare. Section 6 concludes.

2. Experiment

2.1 Sample

The sample consists of 160 people aged between 18 and 85 years. We conducted the experiment

in Paris, France, in multiple one-hour sessions in December 2009. The sample of participants

was randomly selected based on the quota method. Participants were contacted by phone and

informed that they will have to reply to questions about food during one hour with a 15-euro

participation fee. The sample is relatively representative of the age-groups and the socio-

economic status of the population of the city although retired people are slightly over-

represented.

In our experiment, the sample is divided into four groups (see the explanation and the

figure 1 below) and participants are randomly assigned to one group. A Pearson chi-square test

1 One exception is Gifford and Bernard (2004).

5

shows that the four groups are not significantly different from each other in terms of participants’

socio-economic characteristics (gender, age, education, income, household composition).2

2.2 Product

The experiment focuses on a 100g plastic package of farmed, midsize, shelled, cooked and

refrigerated shrimps. Cooked and refrigerated shrimps are the most consumed shrimps in France

(two-thirds of all consumption of shrimps both in value and quantity in 2008 according to the

FranceAgriMer, 2008).3 As no major brand dominates the market, the private brand (linked to a

French supermarket) is concealed to avoid any influence of this supermarket brand.

For fair trade and environmentally friendly shrimps, we add the corresponding label to

the picture of the regular product (instead of using real labeled products). Indeed for really

measuring the marginal value of the environmental and fair trade characteristics with sequential

choices, regular, environmentally friendly and fair trade varieties would have to be similar in a

maximum number of elements, namely, brand, sauce, weight, packaging, and price. Such a

similarity between regular and environmentally friendly varieties did not exist and no fair trade

shrimps were sold on the French market at the time of the study.4 In this last case, the experiment

allows us to measure consumers’ WTP for the hypothetical fair trade variety and evaluate the

possible benefits linked to the introduction of a fair trade label, which is likely to appear in a near

future for shrimps. Participants are however not informed during the experiment that the fair

trade variety is not currently available on the market.

2 The socio-economic characteristics of participants within each group and the results of the Pearson chi-square test

are available upon request.

3 Statistics do not distinguish between shelled and non-shelled shrimps.

4 Furthermore, the cold process linked to refrigeration makes the sale/distribution of real products to participants

hazardous in terms of food safety.

6

Based on the previous literature, the risks of possible hypothetical biases can be

downplayed regarding the welfare measures, since the marginal WTP (namely the difference

between WTP expressed under different choices) is used for computing the value of information

that is the surplus variation coming from revealed information (see also footnote 5). By

comparing hypothetical and non-hypothetical responses, Lusk and Schroeder (2004) show that

marginal WTP for a change in quality/characteristic is, in general, not statistically different

across hypothetical and real payment settings. Moreover, Camerer and Hogarth (1999) show that

performance-based financial incentives have little effect on mean responses.

2.3 Experimental design and revealed information

The experiment is divided into several stages as described in figure 1. Participants receive

general instructions and sign a consent form. They fill in an entry questionnaire on consumption

behavior and socio-demographic characteristics. Five successive rounds of WTP elicitation are

then organized. For each round, we first provide some information to the participants. After the

information revelation, they fill in a multi-price list (or payment card) presented on a paper sheet,

which allows elicitation of the WTP. Finally, participants complete an exit questionnaire and

receive the €15 compensation.

Insert figure 1 here

While the complete information revealed to participants is given in appendix, it is

possible to sum up the content delivered at different points in the experiment as follows:

- In the first round, we provide basic information about the shrimp, including the range

of existing prices observed in supermarkets (between €1.50 and €4).

- In the second round, we provide brief information on concerns about the environment

and working conditions.

7

- In the third and fourth rounds, we reveal additional information about possible

environmental and working conditions attributes linked to shrimps. We divide our

sample into four groups and we vary the type (positive vs. negative) and the order of

information (environmental vs. social) provided to each group. The two first groups

(groups ironmentSocial_Env

PositiveInfo and

Social_tEnvironmen

PositiveInfo ) receive positive information, while the

two last groups (groups ironmentSocial_Env

NegativeInfo and

Social_tEnvironmen

NegativeInfo ) receive negative

information. Information on working conditions is delivered before information on

environmental conditions to groups ironmentSocial_Env

PositiveInfo and

ironmentSocial_Env

NegativeInfo , while

groups _SocialtEnvironmen

PositiveInfo and

Social_tEnvironmen

NegativeInfo receive information on environmental

conditions before information on working conditions.

- Finally, in the fifth round, we reveal information about safety. Previous experiments

show that safety information is a priority for participants eclipsing other

characteristics (Marette et al., 2009), while in this experiment we want to ignore food

safety considerations except at the end of the experiment.

A multiple price list is presented on a sheet of paper to elicit participants’ WTP. During

each choice phase, participants are asked to choose whether or not they will buy the product for

prices varying from €0.25 to €4 with a 25-cent interval between possible choices. A color picture

of the shrimp package is posted on the paper sheet. For fair trade and environmentally friendly

shrimps, we also post a “fair trade” or “environmentally friendly” label. For each price,

participants have to check off either “yes”, “no” or “maybe” denoting their purchase preferences.

The option “maybe” is useful for capturing hesitation that differs from a firm “yes”. For each

choice #i with i=1,...,5, the WTP is determined by taking the highest price linked to a “yes”

choice. If no “yes” is checked off, we set the WTP to zero. If “yes” is always selected, we set the

WTP to €4.

8

Andersen et al. (2006) underline two disadvantages of the multiple price list. The first

disadvantage is the interval response eliciting interval from participants rather than point

estimates for WTP. With our experiment, the 25-cent interval offers enough precision for the

elicited WTP. The other disadvantage is the framing effect with a psychological bias towards the

middle of the multiple price list for choices made by participants. Andersen et al. (2006) control

for this effect by changing the boundaries of the multiple price list. In this paper we do not

change the boundaries of the list and therefore do not control for this framing effect, since we

focus on the impact of the revelation of information and messages. The psychological bias is

plausible for the first round of our experiment, 18.1% of participants expressing a WTP of €2.

However, this effect disappears after the revelation of information. Only 11.3% of participants

make a bid of €2 in choice #2, and this percentage becomes less than 6% in choices #3, #4 and

#5. Despite these limitations, the multiple price list methodology is useful for providing

information regarding the consumers’ WTP.

3. Experimental results

Figure 2 shows the average WTP in euro for 100g of shrimps. This average takes into account

bids by all participants, including the ones with WTP equal to zero. The standard deviation is

reported in parentheses. Recall that the two upper groups ironmentSocial_Env

PositiveInfo and

Social_tEnvironmen

PositiveInfo

(respectively the two lower groups ironmentSocial_Env

NegativeInfo and

Social_tEnvironmen

NegativeInfo ) receive positive

information with labels (respectively negative information). The x-axis of each graph details

under each bar the round of choice i with i=1,...,5 and information preceding the choice leading

to the WTP elicitation. The indicators Δ isolate the significant impact of a single round of

additional information. We test for the significance of the WTP differences following a single

round of information (namely, between WTP #i and WTP #i+1) by using the Wilcoxon test for

paired samples and indicate the significant differences at the 1 %, 5%, and 10% level.

9

Insert figure 2 here

Seven main results could be highlighted:

(i) The initial WTP before the revelation of any information about social, environmental

and health-related conditions is similar across groups. A Kruskal-Wallis test concludes that the

valuation for the first round across the four groups is not statistically significant.

(ii) The short and vague message we provide does not influence consumers with pre-

existing knowledge about the question studied in the experiment. A Kruskal-Wallis test

concludes that the valuation for the second round is similar across groups. Furthermore, the

revelation of short and general information about environment and social conditions before

choice #2 does not lead to a significant change in WTP for three of the four groups. For the last

group (group ironmentSocial_Env

NegativeInfo ), the Wilcoxon test for paired samples shows that the difference

in WTP between the first and the second bars is significant but only at the 10% level. This result

differs from conclusions showing that short and simple information is efficient for changing

WTP (Wansink et al., 2004). Clearly, there are no definitive conclusions that depend on both the

products and the characteristics at stake.

(iii) A sufficient level of precision in the revealed information is efficient for changing

WTP. Participants react to the second and more precise round of information (before choice #3).

The differences in WTP between the second and third bars of each graph are significant, except

for one group (group Social_tEnvironmen

NegativeInfo , significant at 10.1% only).

(iv) Positive information has a larger impact on consumers’ WTP in absolute value than

negative information. The average variation in absolute value between the second and third bars

is equal to €0.73 and to €0.75 for the two groups receiving positive information and only to

€0.34 and to €0.48 for the two groups receiving negative information.5 This diverges from some

5 For positive information the premium equal to €0.75 corresponds to an increase of 35% of the WTP. Some farms

in Madagascar recently offer organic shrimps with a price premium of 25% compared to regular products sold in

10

results on food safety showing that negative information has a larger impact than positive

information (Fox et al., 2002; Hayes et al., 1995). One explanation for our result could be that

consumers tend to undervalue the distinctive advantage of the label and give higher ratings to

products with a label and additional details covered by that label (Hoogland et al., 2007).

(v) Different labels have similar effect. The Mann-Whitney-U test across the two groups

ironmentSocial_Env

PositiveInfo and

Social_tEnvironmen

PositiveInfo reveals that valuations for the third rounds are not

statistically different, even if labels are different. Interestingly, the same result occurs when one

focuses on the two groups receiving negative information. The fact that no label dominates other

labels may explain the appearance of new labels on the market, since some producers may

benefit from higher consumers WTP with a new signal that is different from other labels. This

result also suggests that farmers from developing countries could consider different quality

improvements associated with different labels to capture consumer WTP.

(vi) Consumers’ WTP does not seem to be influenced by different labels or negative

information. The additional information about the second characteristic (before choice #4 leading

to the 4th bar) does not impact WTP for the four groups of participants. This is true whatever the

type of information (positive vs. negative) and in case of positive information whatever the type

of labels (environmentally friendly or fair trade). Similarly, the last information about safety

(before choice #5 leading to the 5th bar) clearly matters for the two groups, which previously

received only positive information. The effect is much smaller and even not significant for the

two groups previously receiving negative information.

other farms (see Hervieu, 2009). This premium of 25% at the farm gate is a lower bound since middlemen and

supermarkets tend to increase it when offering the organic product. This simple comparison suggests that our

welfare measures do not suffer from a hypothetical bias

11

(vii) Experiment results are driven by neither an information effect per se nor a tiredness

effect. First, the revelation of short and general information before choice #2 does not lead to a

significant change in WTP for three of the four groups, suggesting the absence of an information

perturbation whatever its content (information effect per se). As a consequence, the significant

impact of information before choice #3 can really be interpreted as a shock coming from the

content of positive information or negative information. Second, the significant impact of the last

round of information before choice #5 indicates that participants’ attention does not suffer from a

tiredness effect.6 The non-significant impact of information before choice #4 can therefore be

interpreted as the absence of an additional premium for a second characteristic (under both

positive and negative information).

4. Econometric estimations

We now provide more explanations regarding the results of figure 2 and investigate the

determinants of changes in WTP. To do so, we regress difference in WTP expressed by

participants between choices #i+1 and #i (with i=1,…,4) on information and participants’

characteristics. Results are presented in table 1. Since participants provide several answers, our

sample is a panel and we could use a panel estimator. However, each participant makes multiple

choices, and therefore there could be some correlation across data points relating to WTP. To

deal with this issue, we use the random effects panel estimator.7

We first examine whether the revelation of positive information (respectively negative)

before choice #i+1 increases (respectively decreases) participants’ WTP (column 1). We assume

6 Due to tiredness, participants’ attention could decrease during the experiment and they react less (or even do not

react) to new information. 7 We also tested for the influence of participants’ socio-economic characteristics (sex, age, presence of children in

the household, and level of education). Because none of the estimated coefficients on these variables was

statistically significant, we decided to remove them from the estimations.

12

that three types of information are revealed during the experiment: (i) neutral information (i.e.

general and brief messages #1 and 2 provided to all participants), (ii) positive information

(messages #3 and 4 delivered to groups ironmentSocial_Env

PositiveInfo and

Social_tEnvironmen

PositiveInfo ), and (iii) negative

information (messages #3 and #4 delivered to groups ironmentSocial_Env

NegativeInfo and

Social_tEnvironmen

NegativeInfo , and

message #5 to all groups). To test the impact of information on participants WTP, we therefore

define two dummies: one for positive information and one for negative information. The first

dummy (respectively the second one) is set to one if positive information (respectively negative

information) is revealed and 0 otherwise. Results suggest that revealed positive information

increases participants’ WTP, while negative information decreases it. Estimated coefficients on

both dummy variables are significant at the 1% level.

In column (2), we interact the dummies on positive and negative information with

dummies on the type of revealed information before choice #i+1: environmental, social or

health-related. Interestingly, only three interaction terms are significant at the 1% level and have

the expected sign; positive social and positive environmental information increases WTP, while

negative safety information reduces WTP. The two other interaction terms (namely, negative

social and negative environmental information) are not significant. Furthermore, the F-test shows

that the interaction terms (positive x social information and positive x environmental

information) are not significantly different.

Column (3) examines if the order in which information is presented impacts the WTP, i.e.

if information about social conditions (respectively environment) has a similar or different effect

depending on whether it is presented before or after environmental concerns (respectively social

conditions). To perform this analysis we define two dummies: “first information” equals to one

for information provided before choice #3 (0 otherwise) and “second information” set to one for

information revealed before choice #4 (0 otherwise). We then interact these two dummies with

the previous dummies on positive/negative and social/environmental information. Results show

13

that only the positive message on social or environmental conditions delivered before choice #3

significantly affects the differences in WTP. The first negative message before choice #3 and the

second (positive or negative) message before choice #4 have no significant influence. Besides,

the F-tests suggest that the interaction terms (positive x social x first information and positive x

environmental x first information) are not significantly different. These results are in line with

figure 2 and confirm that consumers’ valuations are more sensitive to positive information than

to negative information.

Finally, in columns (4) and (5), we investigate whether information affects differently

participants belonging to alternative demographic groups. We interact information variables with

two socio-economic characteristics of participants: sex and presence vs. absence of children in

the household. Previous results remain qualitatively unchanged. The estimated coefficients on

these interaction terms show some differences between male and female and between

participants living with/without children at home. However, the F-tests suggest that these

differences are not statistically significant. This absence of influence of socio-economic

characteristics is often found in experimental economics, especially when one controls for the

individual effect (i.e. for the correlation across the multiple choices made by each participant) as

we do. It suggests that reactions to information seem similar across the sample of participants

and, by extrapolation, for the overall French population. In other words, reactions are relatively

similar whatever the people attending the experiment.

Insert table 1 here

5. Value of information and consumer welfare

The WTP can be used to determine the consumer surplus and the value of information.

Following Foster and Just (1989) and Teisl et al. (2001), information is welfare enhancing if

consumers change their consumption behavior. The contribution of our experiment is twofold.

14

First, we investigate whether positive and negative information for the same characteristics has a

similar impact on participants’ behaviors. Second, we combine both positive and negative

information to estimate consumer welfare. Our approach therefore differs from the previous

studies focusing only on the effect or/and the value of information related to the labeled variety

(Huffman et al., 2003, 2007; Lusk et al., 2005; Rousu et al., 2004, 2007; Rousu and Lusk, 2009).

Consumers’ knowledge is affected by various parameters including levels of consumer

attention, advertising, ambient opinion, media coverage, regulation, etc. There are many contexts

of information perception that makes difficult the precise characterization of “correct” or

“incorrect” information. Some consumers (or nations) pay more attention to positive information

and some others are more concerned by negative information. Many limits in consumption are

coming from difficulties to collect, understand and/or process the load of information (Ariely,

2000). Imperfect recall, lack of time before purchasing or/and confusion about complex

information characterize many consumers in the supermarket and potentially bias their

consumption decisions. Risks of consumers’ confusion and difficulties to understand complex

recommendations with positive and negative points of views diminish the information efficiency

(Sasaki et al., 2011).

All these facts explained why we describe different contexts of information. Surplus

analysis under different contexts could help streamline debates around food media coverage and

improve regulatory efficiency regarding the revelation of information. Both positive and negative

messages are complementary since they concern different varieties in competition. Positive

information is clearly related to the variety with the label (organic or fair trade), while negative

information is related to the regular variety without label. Revealing information about the

environmentally friendly or fair trade variety does not mean that all information is revealed about

the ‘harmful’ one and the extent of the environmental and social problems. The consumers’

perception of the regular variety is also modified by the revelation of information about the

15

regular variety.8 This revelation impacts the intensity of the perceived variety differentiation as

precisely described in section 5.3”

We focus on information delivered just before choice #3. This information significantly

impacts WTP for three of the four groups (see figure 2)9, and therefore its welfare impact can be

computed by comparing 3WTP and 1WTP . The focus on this third message presents three main

advantages:

(i) When positive information linked to a label is revealed, consumers are aware of a

new purchasing option. In this case, we may simulate the introduction of a new variety of the

product signaled by a label (fair trade or environmentally friendly label).

(ii) When negative information about the regular variety is revealed, consumers may

take different action compared to the absence of precise information. In this case, we may

determine a non-internalized damage.

(iii) Eventually, we combine both types of positive and negative information to have a

complete view when different varieties of a given product coexist on the market. The

combination of detailed and complete information about each variety provides a balanced

estimation of the value of information. In our context, each group provides well-informed,

thoughtful preferences about one variety without interferences with competing varieties since the

consumers’ focalization is guaranteed.

8 Negative information on the regular product would not be revealed by the regular industry via advertising. It may

be revealed by intense campaigns made by environmentalist groups that can be highly vocal like Greenpeace. Rousu

et al. (2004) detail a method for valuing information as correct despite the eyes of the entity providing it. In this

paper, we do not focus on a specific entity, but rather insist on different sources of information linked to different

products.

9 For group ‘Neg_env/soc’, information is significant at 10.1% only when we compare WTP2 and WTP3. However,

the comparison of WTP1 and WTP3 for this group shows a statistically significant difference at 1%.

16

5.1 Positive information about the new (labeled) variety

With positive information linked to a labeled variety for groups ironmentSocial_Env

PositiveInfo and

Social_tEnvironmen

PositiveInfo , we can measure the impact of a label introduction on participants’ surplus. This

label leads to the introduction of a new variety complying with the label requirements. Before

measuring the value of information, we determine the participants’ purchase choices in periods 1

and 3. We assume that a participant purchases a good if his WTP for the good is higher than the

price observed on average in the supermarkets in France. Before choice #1, only “regular”

shrimps are offered, and the participant can choose between two outcomes: regular at price RP

and none. The participant j chooses the option generating the highest utility, namely:

1 1max{ ,0}j jRCS WTP P , (1)

where the subscript 1 denotes the bid linked to choice #1 for a participant j (with j=1,…,N).

When a label is introduced at price LP (before choice #3), the participant can choose

between three outcomes: regular variety, labeled variety, and none. She/he chooses the

alternative, which generates the highest utility, and thus:

3 1 3max{ , ,0}j j jR LCS WTP P WTP P . (2)

We now turn to the value of information by using two metrics to compute the average

value for each group ironmentSocial_Env

PositiveInfo or

Social_tEnvironmen

PositiveInfo receiving positive information. The

participants’ surplus change from a label introduction, if all participants are fully informed about

the label, is:

3 1 1 3 11 1

[ ] [max{ , ,0} max{ ,0}]K K

j j j j jR L R

j jKLabel

CS CS WTP P WTP P WTP P

CSK K

. (3)

where K can take two values. First K=SP, where SP is the number of switchers receiving positive

information who start to consume the labeled variety after its introduction. In this case, the

information revelation only modifies the surplus of these switchers. Conversely, the surplus

17

variation is zero for participants who do not change their behavior, namely by continuing to

choose the regular variety or nothing when the label is introduced and when RP is constant

(which is the case under constant-return to scale for producers). Second, K=NP, where NP is the

overall number of participants within a group receiving positive information. The measure given

by (3) is similar to the one provided in papers focusing on the introduction of genetically

modified products (Huffman et al., 2003, 2007; Lusk et al., 2005).

5.2 Negative information about the regular variety

With negative information revealed to groups ironmentSocial_Env

NegativeInfo and

Social_tEnvironmen

NegativeInfo , we can

measure the non-internalized damage10 linked to the lack of precise information (as before

choice #1). When the negative and precise information (before choice #3) is revealed to these

two groups, some participants stop buying the product. For a participant j, the non-internalized

damage linked to the absence of complete information about a characteristic before choice #1 is

][ 31jj

jj WTPWTPID where Ij is an indicator variable taking the value of 1 if participant j is

predicted to have chosen the regular variety at price PR with 1j

RWTP P in choice #1 (and 0

otherwise). The non-internalized damage reduces the participant’s surplus associated with choice

#1. By taking into account 1 1max{ ,0}j jRCS WTP P (see equation 1), the overall surplus is:

][0,max 31111jj

jRjjjj WTPWTPIPWTPDCSCSC (4)

When negative information is revealed, outcomes remain unchanged (regular variety or

none). However, participants may adjust their consumption, and the surplus becomes

3 3max{ ,0}j j

RCS WTP P . For each group ironmentSocial_Env

NegativeInfo or

Social_tEnvironmen

NegativeInfo , the value of

10 This non-internalized damage is slightly different from the cost of ignorance suggested by Foster and Just (1989).

In Foster and Just’s (1989) framework, consumers incur a cost of ignorance from consuming a contaminated product

that could cause detrimental health effects without knowledge of the adverse information.

18

information or total surplus change for participants linked to the revealed information is given

by:

3 11

3 1 1 31

[ ]

[max{ ,0} max{ ,0} [ ] ]

K j j

jKNeg

Kj j j j

R R jj

CS CSC

CSK

WTP P WTP P I WTP WTP

K

(5)

where K=SN (the number of switchers) or K=NN (the overall number of participants within a

group receiving negative information). In this context, SN is the number of switchers that stop

consuming the product when negative information is revealed before choice #3.

5.3 Combination of positive and negative information

Eventually, one can combine negative and positive information across the groups by integrating

the average values of the non-internalized damages coming from groups receiving negative

information in the surplus of the groups receiving positive information. This combination is

likely to be more satisfying compared to previous situations presented in subsections 5.1

(revelation of positive information on the labeled variety only) and 5.2 (revelation of negative

information on regular variety only) as it equally balances the weight of both groups for positive

and negative information. The value of information is computed to measure the impact of the

new variety linked to a fair trade or environmentally friendly label by including the average

value of the non-internalized damage linked the purchase/consumption of the regular variety.

The existence of a new label may also lead many consumers to question the existence of “bad”

characteristics linked to the regular variety.

For the two groups receiving negative information, the average value of the non-

internalized damage linked to the regular variety is:

19

N

N

N

jj

N

j

jjj

I

WTPWTPI

DE

1

131 ][

)( (6)

where 1

NN

jj

I is the number of participants who purchase the good based on WTP revealed by

choice #1.

The average measure E(D) coming from group ironmentSocial_Env

NegativeInfo is integrated in the

participant’s j surplus for group ironmentSocial_Env

PositiveInfo receiving positive information (the equivalent

can be made for groups Social_tEnvironmen

NegativeInfo and

Social_tEnvironmen

PositiveInfo ). By taking into account

1 1max{ ,0}j jRCS WTP P given by (1) and linked to the choice #1, the overall surplus with only

the regular variety on the market is:

)(0,max)( 111 DEIPWTPDECSCSD jRjjj (7)

The introduction of the label leads to two situations, namely case 1 and case 2. Under

case 1, participants faced with a label are only aware of positive information coming from the

label for choosing between options (regular, label, and none), but their choices may be distorted

because of a lack of information regarding the regular variety. Their surplus is

)(0,,max 313 DEIIPWTPPWTPCSD RjL

jR

jj , where IIjR is an indicator variable taking the

value of 1 if participant j is predicted to have chosen the regular variety at price PR when the

label exists on the market (and 0 otherwise). In this case, the value of information is:

K

DEIPWTPDEIIPWTPPWTP

K

CSDCSD

CSD

K

jjR

j

jR

L

j

R

j

K

j

jj

LabelK

1131

113

)(0,max)(0,,max (8)

20

where K=SP (the number of switchers) or K=NP (the overall number of participants within a

group receiving positive information).

Under case 2, participants faced with a label are fully aware of both positive information

coming from the label and negative aspects linked to the regular variety when they choose

between regular, label or none. The emergence of a label may indeed lead to awareness of

negative aspects linked to the regular variety. For instance, participants realize to what extent the

production of the regular variety pollutes the environment or does not provide decent working

conditions. Thus, their surplus is 0,,)(max 313 Lj

Rjj PWTPPDEWTPCSE . Their WTP for

the regular variety at price PR is )(1 DEWTP j if they are aware of the negative aspect based on

the average non-internalized damage E(D) that is internalized in their demand. In this case, the

value of information is:

K

DEIPWTPPWTPPDEWTP

K

CSDCSE

CSE

K

jjR

j

L

j

R

j

K

j

jj

LabelK

1131

113

)(0,max0,,)(max

(9)

where K=SP (the number of switchers) or K=NP (the overall number of participants within a

group).

All these measures indicating the value of information (or the welfare variation linked to

more complete information) lead to the following results presented in table 2. Based on

observations across several supermarkets, we use PR = €2.2 and a price premium equal to 25%

for the varieties with a label leading to PL = €2.75 (Hervieu, 2009).11

11 Hervieu (2009) emphasizes that environmentally friendly shrimp production requires 25% more work than regular

shrimps, which leads us to assume a price increase equal to 25%. Since fair trade shrimps are not available in

France, we assume that the price premium for such shrimps is similar to the one observed for environmentally

friendly shrimps.

21

The first three lines of table 2 measure the value of information as in previous papers by

focusing on positive information related to the labeled variety. Lines 4-6 measure the value of

negative information linked to the regular variety. The welfare variation linked to the revelation

of negative information is higher than the one observed when positive information is revealed.

This result comes from the higher number of switchers stopping consumption of the regular

variety when negative information is revealed. It shows the importance of taking into account the

non-internalized damage coming from the regular variety when measuring the value of

information.

Results based on the combination of positive and negative information are presented in

the bottom part of table 2. Recall that negative information comes from the average damage E(D)

for which participants are aware or not. Case 2 corresponds to the situation of perfect

information. Both cases 1 and 2 show a relative large value of information, which is higher than

that observed when only positive information related to the labeled variety is accounted for.

Thus, by focusing only on information related to the new (labeled) variety, previous welfare

measures underestimate the value of information. Robust evaluations need to combine both

negative and positive information. Interestingly, our results also show that the two average

values of information over all participants given in the last line of table 2 are relatively high

compared to a price PR = €2.2 considered for the estimations of the surplus variation coming

from information. This suggests a high-social benefit linked to the complete revelation of

information to consumers. This benefit should be compared to regulatory costs coming from

quality monitoring and advertising efforts.

Insert table 2 here

Moreover, fully revealing negative and positive information generates higher WTP by

developed countries’ consumers that can be of benefit to producers from developing countries.

The results of the experiment could therefore be used to measure the increase in producers’

marginal income from a complete development of labels, where both labeled and regular

22

varieties would be offered everywhere in France. The simulation is as follows: In France, the

consumption of farmed tropical shrimps is 56,548,800 kg per year.12 To carry out the income

estimation, we consider the participants who switch to labeled varieties after the information

revelation. Among them, we distinguish the percentage x of participants who purchase the

regular variety before the label introduction and the percentage y of participants who do not

purchase shrimps before the label introduction. The income increase for producers comes from

these consumers changing varieties at a better price PL compared to PR and from the new

consumers purchasing goods at price PL. Ignoring any cost consideration the increase in

producers’ marginal income is defined by 56,548,800*[x*(PL - PR) + y* PL] and presented in

table 3. Negative information is not taken into account since this information is not widely

broadcasted compared to the labels. Although these estimations are imperfect, they show

significant increase in producers’ income that may cover additional label-sunk costs (not passed

into the price PL) and generate higher income compared to the existing situation.

Insert table 3 here

6. Conclusion

These results suggest that environmental and social labels have similar effects on participant

willingness-to-pay when it is first presented. However, we also highlight a fast-diminishing

interest in the second-presented characteristic. These results suggest (i) the importance of the

first/major information sent by advertising campaigns to consumers to signal new varieties, (ii)

the difficulties of developing added attributes for new varieties when one attribute dominates a

market and (iii) the possibilities of cumulating several labels to attract consumers with different

12 This volume is calculated by taking the total consumption of shrimps in France in 2008 and multiplying it by the

percentage of shrimps that are tropical (80%) and by the percentage of farmed shrimps (60%). Statistics sources:

FranceAgriMer (2008) and United Nations Food and Agriculture Organization (2009).

23

priorities regarding the conditions of production and not consumers with utilities favoring added

attributes. It also means that farmers from developing countries should carefully consider the

type of labels they should use to improve quality and capture consumer WTP.

Using welfare variation from the environmentally friendly or fair trade labels, we also

show that the absence of negative information linked to the regular variety may lead to an

underestimation of the value of information associated with the label. The consumers’ choices

are distorted since negative information linked to the regular variety is not internalized by

consumers. The segment for products with labels could be expanded. This last result underlines

that the benefit of complete information in terms of welfare and profit variations could be high

not only for French consumers but also for shrimp producers developing high-quality products.

Improving the quality of agricultural products is one way to increase income in developing

countries, even if choosing the type of labels or the advertising strategy is challenging for these

producers.

Our application is simple and based on various assumptions. In real situations,

participants are limited in their ability to collect complete information about products, and they

may be confused about different labels. Moreover, the cost of building the reputation of a new

label indicating environmentally friendly and/or fair trade products needs to be taken into

account. Some extensions also could be considered for completing the results of this paper. An

experiment in some shrimp farms in Asian countries to find out producer willingness to invest in

high-quality/labeled products may suitably complete the analysis. A second extension of this

work could consist in investigating the psychological mechanisms that drive the WTP expressed

by consumers. Do environmental and social concerns appeal to the same type of motivations?

How do consumers perceive negative and positive information when it comes to ethical

consumption? A third extension could consist in deeply studying what information should be

delivered to consumers to consider them “well-informed” with thoughtful preferences. It is likely

to include both positive and negative information sources. A new experiment could then be used

24

for computing the value of not receiving information, of receiving positive information before

negative information, and of receiving negative information before positive information.

7. References

Andersen, S., Harrison, G.W., Lau, M.I., Rutström, E.E., 2006. Elicitation Using Multiple Price

List Formats. Experimental Economics 9(4), 383-405.

Ariely, D., 2000. Controlling the Information Flow: Effects on Consumers’ Decision Making and

Preferences. Journal of Consumer Research 27(2), 233-248.

Arnot, C., Boxall, P.C., Cash, S.B., 2006. Do Ethical Consumers Care About Price? A Revealed

Preference Analysis of Fair Trade Coffee Purchases. Canadian Journal of Agricultural

Economics 54(4), 555-565.

Bernard, J.C., Bernard, D.J., 2009. What Is It About Organic Milk? An Experimental Analysis.

American Journal of Agricultural Economics 91(3), 826-836.

Blend, J.R., van Ravenswaay, E.O., 1999. Measuring Consumer Demand for Ecolabeled Apples.

American Journal of Agricultural Economics 81(5), 1072-1077.

Bougherara, D., Combris, P., 2009. Eco-labelled Food Products: What are Consumers Paying

for? European Review of Agricultural Economics 36(3), 321-341.

Camerer, C., Hogarth, R., 1999. The Effects of Financial Incentives in Experiments: A Review

and Capital-Labor-Production Framework. Journal of Risk and Uncertainty 19(1-3), 7-42.

Foster, W., Just, R.E., 1989. Measuring Welfare Effects of Product Contamination with

Consumer Uncertainty. Journal of Environmental Economics and Management 17(3), 266-

283.

Fox, J.A., Hayes, D.J., Shogren, J.F., 2002. Consumer Preferences for Food Irradiation: How

Favorable and Unfavorable Descriptions Affect Preferences for Irradiated Pork in

Experimental Auctions. Journal of Risk and Uncertainty 24(1), 75-95.

25

FranceAgriMer, 2008. Bilan Annuel 2008 – Consommation des Produits de la Pêche et de

l’Aquaculture. Montreuil-sous-Bois: Direction Marchés, Etudes et Prospective.

Gifford, K., Bernard, J.C., 2004. The Impact of Message Framing on Organic Food Purchase

Likelihood. Journal of Food Distribution Research 35(3), 19-28.

Hayes, D.J., Shogren, J.F., Youll Shin, S., Kliebenstein, J.B., 1995. Valuing Food Safety in

Experimental Auction Markets. American Journal of Agricultural Economics 77(1), 40-53.

Hervieu, S. 2009. A Madagascar, la Seule Crevette d’Elevage Bio du Monde. Le Monde, April

1st, p.4.

Hoogland, C.T., de Boer, J., Boersema, J.J., 2007. Food and sustainability: Do consumers

recognize, understand and value on-package information on production standards? Appetite

49(1), 47-57.

Huffman, W.E., Rousu, M.C., Shogren, J.F., Tegene, A., 2003. The Public Good Value of

Information from Agribusinesses on Genetically Modified Foods. American Journal of

Agricultural Economics 85(5), 1309-1315.

Huffman, W.E., Rousu, M.C., Shogren, J.F., Tegene, A., 2007. The Effects of Prior Beliefs and

Learning on Consumers’ Acceptance of Genetically Modified Foods. Journal of Economic

Behavior & Organization 63(1), 193-206.

Loureiro, M.L., Lotade, J., 2005. Do Fair Trade and Eco-labels in Coffee Wake up the Consumer

Conscience? Ecological Economics 53(1), 129-138.

Loureiro, M.L., McCluskey, J.J., Mittelhammer, R.C., 2001. Assessing Consumer Preferences

for Organic, Eco-labeled and Regular Apples. Journal of Agricultural and Resource

Economics 26(2), 404-416.

Lusk, J.L., House, L.O., Valli, C., Jaeger, S.R., Moore, M., Morrow, B., Traill, W.B., 2005.

Consumer Welfare Effects of Introducing and Labeling Genetically Modified Food.

Economics Letters 88(3), 382-388.

26

Lusk, J.L., Marette, S., 2010. Welfare Effects of Food Labels and Bans with Alternative

Willingness to Pay Measures. Applied Economic Perspectives and Policy 32(2), 319-337.

Lusk, J.L., Schroeder, T.C., 2004. Are Choice Experiments Incentive Compatible : A Test with

Quality Differentiated Beef Steaks. American Journal of Agricultural Economics 86(2), 467-

482.

Marette, S., Roosen, J., Bieberstein, A., Blanchemanche, S., Vandermoere, F., 2009. Impact of

Environmental, Societal and Health Information on Consumers’ Choices for Nanofood.

Journal of Agricultural & Food Industrial Organization 7(2), Article 11.

Nimon, W., Beghin, J.C., 1999. Are Eco-Labels Valuable? Evidence from the Apparel Industry.

American Journal of Agricultural Economics 81(4), 801-811.

Rousu, M.C., Huffman, W.E., Shogren, J.F., Tegene, A., 2004. Estimating the Public Value of

Conflicting Information: The Case of Genetically Modified Foods. Land Economics 80(1),

125-135.

Rousu, M.C., Huffman, W.E., Shogren, J.F., Tegene, A., 2007. Effects and Value of Verifiable

Information in a Controversial Market: Evidence from Lab Auctions of Genetically Modified

Food. Economic Inquiry 45(3), 409-432.

Rousu, M.C., Lusk, J.L., 2009. Valuing Information on GM Foods in a WTA Market: What

Information is Most Valuable? AgBioForum 12(2), 226-231.

Sasaki, T.D., Becker, V., Janssen, M.A., Neel, R., 2011. Does Greater Product Information

Actually Inform Consumer Decisions? The Relationship between Product Information

Quantity and Diversity of Consumer Decisions. Journal of Economic Psychology 32(3), 391-

398.

Teisl, M.F., Bockstael, N.E., Levy, A., 2001. Measuring the Welfare Effects of Nutrition

Information. American Journal of Agricultural Economics 83(1), 133-149.

United Nations Food and Agriculture Organization, 2009. Fisheries Statistical Collections –

Global Production. Rome.

27

Wansink, B., Sonka, S.T., Hasler, C.M., 2004. Front-label Health Claims: When Less is More.

Food Policy 29(6), 659-667.

Wessells, C.R., Johnston, R.J., Donath, H., 1999. Assessing Consumer Preferences for

Ecolabeled Seafood: The Influence of Species, Certifier, and Household Attributes.

American Journal of Agricultural Economics 81(5), 1084-1089.

World Bank, 2001. Shrimp Farming and the Environment – Can Shrimp Farming Be Undertaken

Sustainably? Washington, DC.

28

Table 1. Influence of information and individual characteristics on WTP differences

Model (1) (2) (3) (4) (5)

Dependent variable Difference in WTP between choice #i+1 and #i

( ij

ij WTPWTP 1 )

Specification Panel random effects

0: male 0: no

children Characteristic

1: female1:

children Positive information (0/1) 0.53a (0.12) Negative information (0/1) -0.29a (0.11) Positive x social info. (0/1) 0.45a (0.15) Positive x social x 1st info. (0/1) 0.97a (0.19) Positive x social x 1st info. (0/1) x characteristic = 0 1.00a 0.90a (0.22) (0.21) Positive x social x 1st info. (0/1) x characteristic = 1 0.93a 1.20a (0.29) (0.34) Positive x social x 2nd info. (0/1) -0.11 (0.19) Positive x social x 2nd info. (0/1) x characteristic = 0 -0.31 -0.07 (0.29) (0.22) Positive x social x 2nd info. (0/1) x characteristic = 1 0.01 -0.20 (0.24) (0.32) Positive x environmental info. (0/1) 0.61a (0.15) Positive. x envir. x 1st info. (0/1) 0.99a (0.19) Positive x envir. x 1st info. (0/1) x characteristic = 0 1.14a 1.10a (0.29) (0.22) Positive x envir. x 1st info. (0/1) x characteristic = 1 0.90a 0.74b (0.24) (0.32) Positive x envir. x 2nd info. (0/1) 0.26 (0.19) Positive x envir. x 2nd info. (0/1) x characteristic = 0 0.29 0.26 (0.22) (0.21) Positive x envir. x 2nd info. (0/1) x characteristic = 1 0.21 0.26 (0.29) (0.34) Negative x social info. (0/1) -0.10 (0.15) Negative x social x 1st info. (0/1) -0.23

29

(0.19) Negative x social x 1st info. (0/1) x characteristic = 0 -0.15 -0.40c (0.26) (0.24) Negative x social x 1st info. (0/1) x characteristic = 1 -0.31 0.04 (0.26) (0.29) Negative x social x 2nd info. (0/1) 0.03 (0.19) Negative x social x 2nd info. (0/1) x characteristic = 0 -0.07 -0.12 (0.32) (0.22) Negative x social x 2nd info. (0/1) x characteristic = 1 0.07 0.39 (0.22) (0.32) Negative x environmental info. (0/1) 0.10 (0.15) Negative. x envir. x 1st info. (0/1) -0.10 (0.19) Negative x envir. x 1st info. (0/1) x characteristic = 0 -0.28 0.02 (0.32) (0.22) Negative x envir. x 1st info. (0/1) x characteristic = 1 -0.02 -0.36 (0.22) (0.32) Negative. x envir. x 2nd info. (0/1) 0.30 (0.19) Negative x envir. x 2nd info. (0/1) x characteristic = 0 0.36 0.29 (0.26) (0.24) Negative x envir. x 2nd info. (0/1) x characteristic = 1 0.24 0.31 (0.26) (0.29) Negative x safety info (0/1) -0.58a -0.58a (0.12) (0.12) Negative x safety info. (0/1) x characteristic = 0 -0.71a -0.55a (0.15) (0.13) Negative x safety info. (0/1) x characteristic = 1 -0.48a -0.65a (0.14) (0.17) Observations 640 640 640 640 640 Pseudo-R² 0.084 0.118 0.162 0.168 0.171

a, b and c: significant at 1%, 5% and 10% respectively. Standard errors in parentheses.

30

Table 2. Value of information and introduction of the labeled variety

Groups ‘Pos_soc/env’ ‘Pos_env/soc’ ‘Neg_soc/env’ ‘Pos_env/soc’

Participants facing the new labeled variety

Participants who switch to the labeled variety after positive information (%)

19% 20.5%

Value of the labeled variety PSLabelCS (switchers) €0.76 €0.53

Value of the labeled variety PNLabelCS (all

participants) €0.14 €0.11

Participants receiving negative information about the existing regular variety

Participants who switch to avoid purchasing the regular variety after negative information (%)

33.3% 30%

Value of information about the regular variety NS

NegCS (switchers) €1.05 €1.45

Value of information about the regular variety NN

NegCS (all participants) €0.35 €0.43

Combination of groups Case 1: Consumers only aware of positive information coming from the label

Average non-internalized damage E(D) €-1.04 €-0.95 Participants who switch to labeled variety after positive information (%)

19% 20.5%

Value of the labeled variety PSLabelCSD

(switchers) €1.27 €1.12

Value of the labeled variety PNLabelCSD (all

participants) €0.24 €0.23

Case 2: Consumers aware of positive and negative information

Average non-internalized damage E(D) €-1.04 €-0.95 Participants who switch to the labeled variety after positive and negative information (%)

45.2% 59%

Value of the labeled variety and information about the regular variety PS

LabelCSE (switchers) €1.03 €0.87

Value of the labeled variety and information about the regular variety PN

LabelCSE (all

participants) €0.47 €0.51

31

Table 3. Income increase for producers (%)a

Information revealed by the label

Positive

& social

information

Positive &

environmental

information

Participants who switch to labeled varieties 19% 20.5%

Purchasing the regular variety before the label

Not purchasing shrimps before the label

9.5%

9.5%

12.8%

7.7%

Increase in producers’ marginal income 14% 11%

a: relative variation (%) compared to the total income for year without any label.

32

a: group ironmentSocial_Env

PositiveInfo includes 42 participants; group

Social_tEnvironmen

PositiveInfo : 39 participants; group

ironmentSocial_Env

NegativeInfo : 39 participants; group

Social_tEnvironmen

NegativeInfo : 40 participants.

Figure 1: Experimental designa

33

Short info.Environ. & socialconditions

Safety risk

Average WTP for shrimpswith labels for group ‘Pos_soc/env’

Average WTP for shrimpswith labels for group ‘Pos_env/soc’

Safety riskFair Trade Environmentallyfriendly

Fair TradeEnvironmentallyfriendly

1.36(1.51)

2.79(1.13)

2.77(1.14)

2.04(1.18)

2.20(1.14)

0

0.5

1

1.5

2

2.5

3

1 2 3 4 5

∆*** ∆***

1.29(1.45)

2.50(1.34)

2.85(1.07)

2.10(1.27)

2.35(1.11)

0

0.5

1

1.5

2

2.5

3

1 2 3 4 5

∆*** ∆***

Short info.Environ. & socialconditions

Negative InfoSocial condit.

Negative InfoEnvironment

Safety risk

Average WTP for shrimpswithout labels for group ‘Neg_soc/env’

Average WTP for shrimpswithout labels for group ‘Neg_env/soc’

Safety riskNegative InfoSocial condit.

Negative InfoEnvironment

Short info.Environ. & social conditions

Short info.Environ. & socialconditions

2.12(1.03)

1.77(1.18)

1.29(1.28)

1.35(1.35)

1.02(1.35)

0

0.5

1

1.5

2

2.5

1 2 3 4 5

∆** ∆*∆*

1.91(1.19)

1.36(1.44)

1.70(1.38)

1.15(1.47)

0.86(1.40)

0

0.5

1

1.5

2

2.5

1 2 3 4 5

Figure 2. Impact of information on WTP (€/100g)a

a: ∆***, ∆** and ∆* denote significant differences at the 1%, 5% and 10% level respectively as tested by the

Wilcoxon test for comparing paired samples.

34

Appendix: Revealed information Initial information before choice #1 “In what follows we will present you information about farmed, midsize, cooked, shelled and refrigerated shrimps. On the market, the average price for 100g of shrimps varies between €1.50 and €4.” General information before choice #2 “There are serious concerns about bad conditions for the environment and workers in shrimp farms from different countries around the world. Many shrimps consumed in France are imported from these countries.” Different types of information before choices #3 and #4 For groups receiving positive information “Fair Trade Shrimps: In some countries, shrimp producers develop fair trade production, with decent working hours, decent wages, and controlled handling of chemical products used by workers to treat and to clean the shrimps in order to protect workers’ health. These practices, on average, significantly increase the production costs. These products are sold with a “fair trade” label in France.” “Environmentally friendly shrimps: In some countries, shrimp producers develop environmentally friendly production schemes. Discharges are limited and pollution is controlled. Furthermore, the quality of water and ecosystems around the farms is preserved. These practices, on average, significantly increase the production costs. These products are sold with a label in France.” For groups receiving negative information “Social concerns: In some countries, among the biggest shrimp producers and exporters, workers in shrimp farms and factories work 12 or 16 hours per day for a very small salary. Furthermore, without any protection, workers handle chemical products that are toxic to human health and used to treat and clean the shrimps. Given the difficulties and the cost of inspection of imported products, it is likely that a large share of shrimps sold in France was produced in such conditions.” “Environmental concerns: Shrimp farms can generate serious environmental problems. In particular, the discharges coming from farms are a source of pollution: deterioration of water quality and of fertility of soils, which were converted into breeding pools. Given the difficulties and the cost of inspection of imported products, it is likely that the production of a large share of shrimps sold in France generated such pollution.”

35

Information before choice #5 “Health concerns: Many bacteriological infections affect shrimp breeding pools. The bad production conditions (bad water quality for example) favor the growth of bacteria. To fight against these bacteria, the shrimp producers use antibiotics and other chemical products that are toxic to human health and therefore forbidden in almost all countries. Given the difficulties and the cost of inspection of imported products, it is likely that some shrimps sold in France were treated with these antibiotics and chemical products toxic to human health.”